The paper doesn't start off well, with its first concrete example being bricklaying. Brian Potter, on his "Construction Physics" blog, has one or more posts discussing the history of attempts to automate that very task.[1] TL;DR: it's not going well, for a variety of reasons.
Most tasks are proving far harder to automate than expected, even the "easy" ones, for task-specific reasons that are obvious to anyone who has closely observed the task in the field. These guys write as though they are desk bound.
Woo-woo like affirmations? Seriously?
How about worker ownership and profit sharing, an improved social safety net, and free retraining?
1. "Where Are The Robotic Bricklayers?" https://constructionphysics.substack.com/p/where-are-the-rob...